WP 96-14
October 1996
Working Paper
Department of Agricultural, Resource, and Managerial Economics
Cornell University, Ithaca, New York 14853-7801 USA
A Unique Measure of the Welfare Effects of Price Support
Programs for Corn on Family-Farm Households
by Size Distribution
by
Tebogo B. Seleka and Harry de Gorter
-
It is the policy of Cornell University actively to support equality
of educational and employment opportunity. No person shall be
denied admission to any educational program or activity or be
denied employment on the basis of any legally prohibited dis­
crimination involving, but not limited to, such factors as race,
color; creed, religion, national or ethnic origin, sex, age or
handicap. The University is committed to the maintenance of
affirmative action programs which will assure the continuation
of such equality of opportunity.
L
A Unique Measure of the Welfare Effects of Price Support Programs for Corn on Family­
Farm Households by Size Distribution
Tebogo B. Seleka and Harry de Gorter
Abstract
On- and off-farm employment offamily farm members has been a permanent phenomenum in U.S.
agriculture. In this paper, a family-farm model is extended to include both on- and off-farm labor
supply decisions for farm-households in the U.S. com sector. A unique empirical measure of
economic welfare is developed to analyze the effects of government price support programs.
Traditional welfare analysis of farm programs ignores on- and off-farm employment decisions by
focusing only on 'producer surplus' at the aggregate sector level. We determine that the appropriate
measure of welfare for the farm-household includes the 'laborer's surplus'. Theoretical results are
derived to show that conventional analysis overstates the benefits of farm price supports because of
the tradeoff between producer and laborer's surpluses. Empirical simulations indicate that the
laborer's surplus is a significant share oftotal farm-household welfare, especially for smaller farm sizes
that comprise the majority of com farmers in the United States. Results also show that the
government programs may have been misdirected if the goal has been to improve the farm income
situation because the few large farms gain much more in aggregate than smaller farms.
-
A Unique Measure of the Welfare Effects of Price Support Programs for Corn on Family­
Farm Households by Size Distribution
Introduction
On- and off-farm employment of family members has been a permanent and important economic
phenomenum in American agriculture (Gardner, 1992; Ahearn and Lee, 1991). Off-farm participation
by households increased from 30 percent in 1929 to 53 percent in 1982 (Ahearn and Lee, 1991). The
proportion of households working 200 or more days per year increased relatively faster than other
categories, from 6.3 percent in 1929 to 34.6 percent in 1982.
Table 1 shows that off-farm income is a significant proportion of total household income. In
1986, an average U.S. farm household earned about $20,212 or 46 percent of total household income
from off-farm sources. Off-farm income was relatively more significant for smaller farm sizes.
Households with annual sales of less than $40,000 generated $22,534 or 96 percent of their total
income from non-farm sources. Large sized households with annual sales of $250,000 and above
generated $17,562 or five percent oftheir income from non-farm sources. The middle income groups
with annual sales of$40,00-$99,99 ($100,000-$249,999) generated $13,780 ($12,602) or 13 percent
(10 percent) oftheir total income from non farm sources. Data for 1989 portrayed a similar picture
(Gardner, 1992).
Table 2 indicates that household labor represents a significant input in com production. The
data shows that owned (unpaid) labor is utilized on the farm more than hired labor. Small sized! offfarm participants (non-participants) spent $1.10 ($2.54) per hectare on hired labor, compared with
$41.09 ($49.35) per hectare on unpaid labor. Medium sized! off-farm participants (non-participants)
spent $3.92 ($3.17) per hectare on hired labor, compared with $26.13 ($29.34) per hectare on unpaid
r-'
labor. Large sized households spent more or less equal amounts on hired and owned labor.
1
Traditional analysis on the social costs of fann price support programs focus only the
'producer surplus' offann-finns at the sectoral level (Nerlove 1958, Wallace 1962, Gardner 1983,
Floyd 1974, Gisser 1993). However, much research has recognized the importance of off-fann
income (Huffinan 1977, Ahearn and Lee 1991, Gardner 1992, Nilsen 1977,) and the need to study
household off-farm labor supply behavior (Sumner 1982, Huffinan and Lange 1989, Lass, Findeis and
Hallberg 1991). Multiple job holding is an integral component of fann-household optimization. The
inclusion ofon- and off-farm labor supply decisions by family fann members requires the use of fann­
household models in analyzing the welfare economic effects ofgovernmental price support programs.
The purpose ofthis paper is to overcome two major limitations of the literature so far on the
welfare economics offann price supports. First, we extend the fann-household model to develop a
unique welfare measure denoted as laborer's surplus (Nakajima, 1986) to include the effects of on­
and off-farm labor supply decisions by family farm members. Own family labor is very significant for
the majority of households both as an input in the farm production process and as an off-fann income
generator. The concept of laborer's surplus is integrated with the conventional producer surplus
analysis of the fann-finn. The implications of the analytical framework developed in this paper are
highlighted by an empirical application to the price support program for com in the United States.
Second, we use this unique framework to evaluate the distributional economic welfare consequences
of government policy across fann size.
The results show that the conventional analysis overstates the benefits of price supports
because of the tradeoff between the conventional producer surplus and the concept of laborer's
surplus integrated into the fann-household model in this paper. Empirical simulations indicate that
laborer surplus is a significant share oftotal farm-household welfare, especially for medium and small
2
­
farm sizes that comprise the majority of com fanns. Sectoral analysis of the current literature masks
the true consequences and distributional benefits from fann programs, especially if the primary
purpose of price support is to solve the 'fann problem' of low fann incomes and rates of return in
agriculture (Gardner, 1992).
The Analytical Model
We employ the basic agricultural household model where a family faces a labor market to
hire-in or hire-out labor at a market wage rate (Singh, Squire and Strauss, 1986, Nakajima, 1965;
1986). Variants of the same model have been developed and employed as a basis to empirically
estimate off-fann labor supply functions for particular U.S. fann households (Huffinan and Lange,
1989; Sumner, 1982; Lass, Findeis, and Hallberg, 1991). The agricultural household model is a
relevant tool of analysis for the economics of fann households since it combines household
production, consumption and labor supply decisions into a single conceptual framework (Huffman,
1991).
Consider a fann household producing a single commodity Q for a given unit price P.
Suppose the fann household faces a competitive labor market in which to hire-in or hire-out labor
at an exogenously given wage rate ofW. The household maximizes utility of money income I and
leisure L, defined as U(I,L). Utility is maximized subject to the fann production constraint Q(r;K),
the time constraint T=L+U and the money income constraint I=V+[pQ-Wr]+wu, where T
represents total time endowment, U denotes total quantity of owned labor devoted to all work
(fanning and off-fann), V is property (exogenous) income, r denotes total fann labor input (both
owned and hired), and K denotes a fixed fann production input (say land).
The objective function and the first order conditions for utility maximization may be
3
-
summarized as l
max [U(I,L) - A(I - V -PQ(r;K) - W(T - L- I))]
1',1.,1,).
PQp(I" ;K) =W
(I a)
I' = V+[PQ(r';K)-Wr']+W(T -L .)
(I b)
U (I' L .)
L
'
=W
(I c)
UI(I' ,L .)
where Qp(r';K) is the marginal product oflabor, UL(I ',L .) denotes the marginal utility of time and
UI(I •,L .) = A is the marginal utility of money income. Since this model consists of a two-staged
recursive process, the household first maximizes farm profits as indicated by condition 1a. This is the
case since L, I and A are not among the arguments of equation 1a. This behavior resembles that of
the farm-firm whereby the sole motive is profit maximization. Equation la is consistent with the
farm-fum's optimization where labor is utilized up to a point where the value of marginal product of
labor is just equated to the market wage rate. In the second stage of the recursive structure, the
household maximizes utility where conditions lb and Ic are simultaneously solved for L' and f,
subject to the solution from condition 1a. 2
1 Note that the farm production constraint and the time constraint were substituted into the
income constraint to yield a single constraint problem.
2 From the first order conditions, equation 1a is solved independently of equations 1band 1c,
meaning that the production side of the model is separable from the consumption side. Optimal
4
­
Figure 1 is the graphical counterpart of the model where the farm-household hires-in labor
and Figure 2 depicts the household hiring-out labor. In panel (a) of Figure 1, the curve UU'
represents the utility function, VvT is the total value product curve and AA' is the money income
constraint. In panel (b), the curve GvMP is the value of marginal product curve and CC' is the
marginal valuation offamily labor curve. 3 Farm profits are maximized at point Q in panel (a) and at
point q in panel (b), where the value of marginal product is just equated to the wage rate. Therefore,
the producer's surplus is measured by the distance DQ in panel (a) and by area PS in panel (b). Utility
is maximized at point L' in panel (a) and point E in panel (b) where conditions Ib and Ic hold (the
marginal valuation offarnily labor CC' equals the market wage rate W). Total wages are W'ELT and
total owned labor cost is ELTC. The laborer's surplus is the residual area LS. Economic surplus is
area PS+LS, which is the sum of the producer surplus and the laborer's surplus (Nakajima, 1986).
The household depicted in Figure 1 enjoys OL units ofleisure and employs Tq' units oflabor on the
farm. Since the time endowment is OT, the household hires-in labor of q'L units for farm use.
Figure 2 is similar to Figure 1, with the exception that the household hires-out labor. Farm
profits are maximized at point Q in panel (a) and point q in panel (b) where the producer's surplus is
recorded at distance DQ or area PS. Utility is maximized at point L' in panel (a) and point E in panel
values from equation Ia are then used in Ib and Ic, to solve for L" and 1". Note that short-run
farm profits (the square bracketed term in condition 1b) are fixed by the time the stage two
solution is determined.
3 At any given point along the curve CC', the marginal valuation offamily labor (MVFL) is
derived by CC'=UdUI . The MVFL curve is not identical to the labor supply curve (Nakajima,
1986). The supply curve of any commodity is independent of the price of the commodity in
question in that it should not shift due to a wage rate change. Instead, a wage rate change should
lead to a movement along a given labor supply curve. Contrary to this, the MVFL curve would
shift in response to a wage rate change. That is, there are numerous MVFL curves, each
corresponding to a particular wage rate (Nakajima, 1986).
5
-
(b) where the laborer's surplus is area LS. 4 Economic surplus is thus measured by the sum ofPS and
LS. The household enjoys OLunits of leisure and utilizes Tq' units oflabor on the farm. Since OT
is the total time endowment, the household commits Tq' units of owned labor on the farm and hiresout Lq' units to off-farm work for wages. Indeed, the producer's surplus measure will understate
actual economic surplus in the case where a farm household is a combination of a laborer's household
and a farm-firm. This will be the case irrespective of whether or not household members participate
in off-farm markets, as evident from Figures 1 and 2.
Economic Effects of a Price Support
We illustrate the importance ofusing a farm household framework in farm policy analysis with
an example of a price support. In Figure 3, the pre-policy farm profits are maximized at point p in
panel (a), where Tfunits ofhousehold labor are devoted to farming. This is also evident in panel (b)
where the value of marginal product of labor hk intersects the wage rate line WW' at point q.
Household utility is maximized at point x, where 01 units ofleisure are consumed. In panel (b), utility
is maximized at point e where the marginal valuation (cost) of family labor sr intersects the wage rate
line WW'. The household devotes Ifunits of time to off-farm work in both panels. The pre-policy
producer's surplus is hqW and the laborer's surplus is W'es. Hence, the pre-policy economic surplus
is hqes.
The introduction of a price support leads to a shift in the total value product curve from vt
to vt' in panel (a). This corresponds to a shift in the marginal value product curve from hk to hk' in
panel (b). Now the household maximizes farm profits at point p' in panel (a) and q' in panel (b) where
Laborer's surplus is based on the sum of owned labor devoted to work (both off-farm and
on-farm). Therefore, even where there is no off-farm work, the laborer's surplus still has to be
measured as in Figure 1.
4
6
-•.
Tf units of household labor are devoted to farming. As a result, household labor devoted to farming
increases by fP units. Farm profits are now higher and the household's money income increases,
leading to a shift in the budget constraint from ag to algI in panel (a). Utility is now maximized at
point x' where it has increased from uu' to vv'. As a result, leisure consumption is increased from 01
to 01' units in both panels. Time devoted to off-farm work declines from M=lfto M'=l'f units. s
The producer's surplus increases by hqq' (from hqW to hq'W), and the laborer's surplus
declines by see' (from Wes to We's).6 Total economic surplus change ofhq'q-see' in panel (b). What
this implies is that an increase in the price of the agricultural commodity causes the household to
increase its producer's surplus and to reduce its laborer's surplus from working (both on and off the
farm). Therefore, when the equilibrium is perturbed by a price support, the marginal benefit from
farming tends to exceed the marginal benefit from working. This causes the household to make a
trade-off between the two surpluses, until the marginal benefit from farming is just equated to the
marginal benefit from working (both on and off the farm). It then follows that the gain in the
producer's surplus must be greater than the loss in the laborer's surplus (hqq'>see'). If this was not
the case, there would indeed be no incentive for the household to make a trade-off between the two
surpluses.
S
See Seleka for complete formal comparative statics results.
6The graph depicts a rotation in the marginal valuation offamily labor curve. This was done
for graphical convenience. In reality, the sr curve would shift, meaning that the new intersection
with the vertical axis would occur at a higher point than point s in Figure 3. Note
however that the vertical distance between sr and srI narrows as the amount of work time is
reduced.
7
­.
Data and Empirical Application
This section presents the data and empirical methodology used in the analysis of the welfare effects
oftarget prices and acreage controls on U.S. com farm-households. The fundamental features of the
u.s. com price support program analyzed in this paper are fixed target prices, deficiency payments,
and acreage limits and set-aside (Gardner, 1992). Acreage set-asides are calculated as a percentage
ofa farmer 'base acreage' (determined by a five-year moving average of actual area planted plus setaside). Deficiency payments are calculated on historical acreage and fixed program yields' (Gisser,
1993). However, a key aspect offarm policy is the voluntary nature of the program whereby some
farmers opt out, given the cost of acreage limits and set-aside.
Most ofthe summary statistics used in this study were obtained from the Economic Research
Service (ERS) of the United States Department of Agriculture (USDA). The data set from which
group level summary statistics were derived was gathered by ERS through the 1991 Farm Cost and
Returns Survey (FCRS). The summary data obtained were: a) group level average com production
cost data per planted acre; b) group level mean values for other pertinent variables; and c) group level
income statements for all farm enterprises. Com farm-households were decomposed into three
categories based upon sales size: small, medium and large. 7 The three sales class categories were
each further subdivided into two groups based upon whether or not individual households participate
in otT-farm work. This yielded a total of six groups of farm households. This classification was
meant to allow for the determination ofhow prevailing com policies affect the respective farm groups
as separate economic agents.
The ERS data is contained in Appendix A. Table A.l provides the per acre production cost
7 Small producers are those with annual sales ofless than $40,000 per farm-household.
Medium producers' annual sales fall within the range $40,000-$249,999 per farm-household.
Large producer are those with annual sales of $250,000 and over per farm-household.
8
-
data for each group's representative household. The variables listed in Table Al are self explanatory,
and are as obtained from
ER~.
Group level mean values of other pertinent data are presented in
Table A2. The variables are listed as obtained with the exception that hours of operator and hours
of other household members were converted from average weekly hours to total hours per year.
Table A3 contains income statement data for the entire farm (for all farm operations). Additional
statistics were obtained from other sources, which we will document along with the discussion of the
model calibration procedure, to which we now tum.
A non-linear algorithm ofGAMS was utilized to solve the U.S. com model. Initially, the
model was calibrated with acreage controls and the target price in place. The production side of the
model, with policy in place, was considered first using Table Al data. Three production factors were
distinguished: unpaid (owned labor); land; and all other inputs. The production function for each
group level representative household belonging to sales class I and participation class j was a Cobby..
(,).
6
Douglas type, Q ij = Aili/X jj '1<ij g, where A is the efficiency parameter,
r
represents labor (only
owned labor for off-farm participants and all labor for non-participants), X denotes all other inputs
(including hired labor for households not participating in off-farm work), K is the quantity ofland,
and y, w, and f> are the respective production elasticities. Unpaid labor was distinguished because
our interest lies in the time allocation mechanism. It is however worth emphasizing that for
households not participating in off-farm work, hired and owned labor were assumed homogenous
whereas they were assumed heterogenous for off-farm participants. Therefore, the production
coefficient y is with respect to all labor in the case of households not participating in off-farm work
9
and with respect to only owned labor in the case of households participating in off-farm work. s
Land was isolated to be able to detennine the effects ofacreage reduction. Other inputs were
aggregated into a single input category. These include non-labor variable inputs, operating capital,
non-land capital, and capital replacement (Table AI).9 Total fixed cash costs (Table AI) were not
included among production inputs. Instead, these expenses were included in the calculation of
exogenous income, as we will discuss later. Cobb Douglas production elasticities were approximated
as factor shares, imposing constant return to scale (CRS).lO CRS was imposed by dividing
expenditure on each factor by total production cost, instead of dividing by the value of output. From
Table A2, we used hours worked by the operator and by other household members and the
respective percentages of hours used in com production, to calculate owned com hours. The wage
rate for each group was then calculated by dividing total value of unpaid labor by annual hours
allocated to com production.
The production side ofthe model was solved by assuming that, initially, com output, the com
The treatment of labor this way was a judgement call. Since those households participating
in off-farm work employ hired labor as well, the time allocation mechanism could not be solved if
we were to assume homogeneity since the model under such an assumption would not allow for
off-farm work and hired labor employment to occur simultaneously. For households not engaged
in off-farm work, the assumption of homogeneity seemed plausible since households hire-in labor
and do not participate in off-farm work.
S
9
As implied earlier, the all inputs category includes hired labor for off-farm participants.
A proof of the relationship between factor shares and production elasticities is contained in
Tyner and Tweeten (1965). Consider the case ofa Cobb-Douglas production technology. The
factor share for input X may be defined as (PxX}/(PQ). A profit maximizing firm would equate
10
the value of marginal product of X to the price of X. Therefore, in this case we can write
aQ/ax = PJP. Multiplying both sides of this equation by XlQ yields the elasticity of production
Ex = (aQ/aX) * (XlQ) = (PxX)/(PQ) =W. This shows that the factor share is equal to the
production function coefficient under a premise that competitive equilibrium reigns and profit
maximization is attained. Along similar lines, Y = (Wr)/(PQ).
10
-
producer price (the target price), the wage rate, the price of other inputs (Px), acreage, and
production elasticities are exogenously given. The price of other inputs was set at unity. Output of
a representative fann was computed by multiplying yield (Table A.l) by com acres (Table A.2). The
target price of2.75 was used as the prevailing producer price of com for 1991 (USDA. 1992).
The pre-policy production side model for a group level representative household contained
three equations
Yij
QIJ.. = A.T..
IJ IJ
x..IJ 1<...IJ6
(,)j'
y··-I
jj
(2a)
(,)
6··
IJ IJ
IJ
IJ
rYij
x..IJ(,)ij-I K·IJ6ij
W .. = Py.. A.,r.Y X .. IJK .. 'J
IJ
IJ
_
Px - Pw..IJ A..IJ IJ..
(2b)
(2c)
where 2a is the production function and 2b and 2c are the two first order conditions for profit
maximization, one with respect to labor (owned labor for off-farm participants and all labor for non­
participants) and the other with respect to all other inputs. The model was solved for optimal factor
demands (r and X) and the production efficiency parameter (A), using a non-linear programming
algorithm of GAMS. Since optimal farm hours of off-farm non-participants reflect both hired and
owned labor, per our assumption of homogeneity, the shares of owned and hired labor in total labor
cost were used to decompose quantities of labor into hired and owned. Short-run profits from com
were derived as total com sales minus the sum of expenditures on all labor and all other inputs
(variable production inputs).
Production results were then used to solve for the consumption and the off-farm labor supply
components. First, hours ofoff-fann work were calculated as annual income from off-farm wage and
11
-
salary income (Table A2) divided by the wage rate detennined above. Total hours of endowment
were calculated by assuming an average number of workers of 1.6 persons per representative
household in each group. This was multiplied by 24 hours per day and 365 days per year to obtain
total time endowment T. Other farm hours (non-com) were approximated by deducting time
allocated to com from total farm time (Table A2). Hours of leisure were derived as total time
endowment minus the sum of off-farm hours, owned com hours, and other farm hours (non-com).
Since households engage in other farm production activities (non-com), as evident from the
income statements (Table A3), earnings from such activities should be considered to better
approximate the household's cash incomes, and total economic surplus. Precisely, restricted farm
profits from non-com activities should be included in calculating money income. Therefore,
Yij
Px = Pu>..IJ A..r.·
IJ IJ
x..IJ(,)jj-l K·IJ6ij
(3)
where R nc denotes short-run profits from non-com fanning activities, G denotes gross cash income
(Table A3), E represents variable cash expenses (Table A3),
TIc- = PQ -
wr- -PxX - is optimal short-
run profit from com (from the production side) and 0 denotes household time spent on non-com
farming activities. Exogenous (property) income (non-wage off-farm income) was computed from
Tables A2 and A3 as the sum of net cash income from off-farm business, net cash income from
another farm and ranch operation, interest and dividends, and other off-farm income (Table A2)
minus the sum of real estate and property taxes, interest and insurance premium (Table A 3).11
As we have noted, fixed cash expenses in Table Al were not included among our various
input categories. Hence, we include them here in computing exogenous income. Data in Table
A3 includes com data of Table Al and as such fixed com costs of Table Al are a part of those
of Table A3.
11
12
-
Money income (I) for each representative household was then calculated as
I..IJ = 1t~.*
+"D.~ +W..U .. +V..
IJ "'iJ
IJ IJ
IJ
(4)
where U=T-L denotes total work time by household members (the sum of time spent on all farming
activities (com and non-com) and off-farm work time). Next, full income (Y) was calculated as the
sum of money income and the value leisure time (Y=I+WT). The Cobb-Douglas utility parameter
with respect to money income was derived as the share of money income in full income ( cc = I/Y).
Similarly, the utility parameter with respect to leisure was computed as the share of the value of
leisure in full income (P = WL * N).12
Next, the market prices of com for the respective groups were determined by diving the gross
value ofproduction by yield. From these, the market price of com for the entire U.S. was computed
12
The utility function for a group level representative household was specified as a Cobb­
a
p.
Douglas type U ij =lij I)L ij
I) ,
where cc and P are the respective parameters. Equations 1band 1c
may be rewritten as
I..+W..L.. = y .. = 1t~.* +W.. T.. +V..
IJ
IJ
IJ
IJ
IJ
IJ
IJ
IJ
Pjj I jj =W..
cc.. L..
IJ
IJ
IJ
Assuming that cc ij +Pij= 1, one can use the fact that Pij = l-ccij in the second equation and
rearrange terms to obtain
cc..IJ (W..IJL..IJ +1..)
=I..IJ
IJ
Using the first equation, one can substitute Yij for the bracketed expression of the above equation
and rearrange terms to obtain ccjj=lijN jj . This implies that the elasticity of the utility function
with respect to money income may be approximated as the share of money income in full income.
Along similar lines, Pfl 1J.. =(W..L..
)N..IJ , which is the share offull income spent on leisure generating
IJ IJ
activities.
13
­
as a weighted average of prices faced by the respective groups. Constant elasticity domestic and
export demand equations were used as approximations of demand equations. 13 The com domestic
and export demand elasticities used were -0.2 and -1.0 (Bullock, 1992). The ratio of com export was
computed using the 1991 export volume of 1,584 million bushels and production of 7,475 million
bushels (USDA, 1993). This ratio was then used to decompose com output (model derived) into
domestic demand and export demand. Using the estimate of the consumer price of com for the U. S.
and the estimates ofdomestic and export demand elasticities and the respective quantities demanded,
the constant elasticity demand shifting parameters were derived.
With the base results determined, we then computed the no policy equilibrium data. Acres
set aside were restored to the sector. This would ideally shift the output supply curve of each
representative farm household to its no policy level. The no policy equilibrium output Q and output
price P were solved endogenously by equating aggregate supply to aggregate demand. The following
set of equations were then simultaneously solved for the no policy output and output price
€ ..
U
QIJ.. = <I> IJ.. p
<1> .. =
IJ
where
E ij
~[
Y
_IJ
y..
..
W
ij
_
J
w·
1J
W
I
P
=(Yij +wil< 1-Yij -wij)
X
I)
l/(l-y-w)
U
1)
6··
AK.. I)
IJIJ
]
(5)
is the price elasticity ofsupply, the first equation is the output supply
The domestic and export demand equations were therefore defined by Q n = 8 nP 1")D and
Q E =8 EP 1")E , respectively. P denotes price, " represents the price elasticity of demand and 8 is
the demand shifting parameter.
13
14
-
function for a group level representative farm (Seleka, 1996), the third expression represents group
level output supply curve, and the last equation equates market supply with market demand. Total
no policy output was then decomposed into domestic and export demand using the demand parameter
estimates from above. Then, no policy variable factor usage, short-run profits and output supply
quantities were computed. We then computed group level and total economic surplus measures for
the no policy scenario (see Appendix B). These were then compared with their base counterparts to
examine the effects of policy on the respective households and the distributional effects across the
individual groups.
Empirical Results
Table 3 presents estimates of Cobb-Douglas production technology and utility parameters.
As indicated, the production technology efficiency parameter estimates (A's) ranged from a low of
2.225 to a high of3.199. Production elasticity coefficients with respect to labor (y 's) ranged from
0.053 to 0.210. The production elasticity coefficient with respect to other inputs (w's) ranged from
0.570 to 0.712, and varied positively with sales class. Production elasticity parameters with respect
to land (l) 's) are within the range 0.216-0.262 and they seemed somewhat invariant across sales
classes, with the exclusion ofthose for the large-participating and the medium-participating category.
The other four groups produced land elasticities of about 0.22 whereas the medium-participating
(large-participating) registered an elasticity coefficient of 0.26 (0.24). Therefore, the data would
suggest that these latter groups contain farms with a relatively higher return on land, compared with
the other four groups whose land coefficients are smaller and somewhat invariant. Table 3 also
presents output supply elasticity estimates ( € 's). As indicated the estimates range from 2.812 though
3.623.
15
-
In general, production elasticity coefficients appear to reinforce the previous findings in the
literature. Floyd (1965) used an estimate of the share of land of 0.2 and that for labor and capital of
0.8. Rosine and Heimberger (1974) estimated the production elasticities for land within the range
0.1079 (1948) and 0.2158 (1970). The production elasticity with respect to labor (hired and owned)
was within the range 0.3918 (1948) and 0.1973 (1970). The land parameter exhibited an upward
trend between 1948 and 1970 whereas that for labor declined during this same period. Based upon
such a trend and the estimates for 1970, one would conclude that the present estimates are consistent
with the results of Rosine and Heimberger (1974). The results of Shumway, Talpaz, and Beattie
(1979) were consistent with the findings of Rosine and Heimberger (1974). Gisser (1993) set the
CES production function coefficient for land at 0.237 and that for all other inputs at 0.763, in his
calibration of the com model (see p.597). Gisser's estimates were adopted from Kawagoe et al.
(1986), who estimated the shares oflabor, machinery, land and fertilizer to be 0.403,0.310,0.237,
and 0.051, for the time period 1930-80.
Utility function parameters are also presented in Table 3. Parameters for money income (a's)
fall within the range 0.185-0.836. These parameters increase with sales class, if we separate off-farm
participants from non-participants. Within any given sales class, participating households tend to
have a higher income share, compared with their non-participating counterparts. A reverse scenario
is true for the coefficients for leisure (p's). Within a given participation class, the parameter for
leisure declines with the increase in farm size (sales class). And within any given sales class, the
leisure coefficient for nonparticipants is higher than that for participants. These results are mainly due
to the fact that small-sized farms have lower money income (hence, a smaller share of money income
in full income) and that non-participating households devote relatively more time to leisure, compared
with their participating counterparts who also devote some of their time to off-farm wage work. We
16
.'
are not aware of any published utility parameter estimates in the literature to compare with the
present estimates.
Most of the coefficients for money income are greater that those for leisure, with the
exception of those for the small-nonparticipating group and the medium-participating group. The
coefficients for money income and leisure are however almost equal at 0.5 for the medium­
participating group. What may cause concern is the data for the small-nonparticipating group, which
registered the leisure coefficient of0.815 and the money income coefficient of 0.185. Is it that these
households value leisure that much or is it because of other factors such as the unavailability of off­
farm work opportunities, which could not be captured by the present model? With the paucity of
data, as is the present situation, there is no telling as to the likely cause of such an occurrence. It is
also noteworthy that, since a single estimate of the number of adults per household was used across
the various classes, the coefficient for leisure will be overestimated for households whose time
endowment is much smaller than the average and will be underestimated for households experiencing
the reverse. This is because we observed work time (farm and off-farm) and then assigned the
residual time to leisure. This procedure was nonetheless the most realistic, given the scarcity of data.
Table 3 also presents estimates of prices. The wage rate varied across groups of farms, and
tended to increase with sales class. That is, the smallest sales class faces lower wage rates, compared
with medium-sized and large-sized classes. The hourly wage rate falls within the range $3.58-$5.14.
The wage rate estimates are lower than published regional and U.S. estimates. For example, USDA
(1992) reports U.S average hourly farm wages for 1991 of$5.62, $5.44, $5.35 and $5.70 for farm
employees working during January 6-12, April 7-13, July 7-13, and October 16-12, respectively. An
attempt to utilize published estimates in the model calibration exercise yielded variable input data that
drastically deviated from observed input usage. Hence, a realistic approach was to utilize the wage
17
­
rate estimates generated from the model. 14
The price of all other inputs was set at unity. The market price of com was estimated by
dividing total com sales by output. As shown in Table 3, the market price of com varied across
categories offarms. Precisely, market prices of com fall within the range $2.27-$2.35 per bushel.
The com market price for the entire U.S. was recorded at $2.31, which was computed as a weighted
average of group level prices. Group level prices fell within the range of published state level
marketing prices. USDA (1993) reported prices falling within the range $2.12-$2.90. The U.S.
average marketing price amounted to $2.37, and was a few cents larger than the present estimate of
$2.31 per bushel. This study has therefore utilized prices that were generated from the summary
statistics afforded to us by the ERS (Table A.l), rather than those published in other (USDA)
sources.
Table 4 presents policy induced changes in endogenous variables and household economic
surplus measures. Let us evaluate the effects of this program sequentially. The results indicate that
the implementation ofthe program under review has caused an increase in variable farm input usage.
The increase in labor F falls within the range 52-216 hours, and it is positively correlated with farm
size. The increase in other inputs X falls within the range 530-9,937 units (dollars), and it is also
positively related to farm size. Hired labor for non-participants increased by 23 (118) hours from
zero hours for small-nonparticipating (medium-nonparticipating) households. These households
witnessed a transition from hiring-out to hiring-in labor. Therefore, these households would have
participated in off-farm wage work in the absence of government intervention.
The large-
nonparticipating class saw an increase in hired labor ofabout 317 hours, following the implementation
14 As we have noted, wage rate estimates were generated by dividing the value of unpaid labor
by the observed quantity of operator and other household labor. Therefore, any alteration of the
wage rates generated will certainly alter hours of owned labor devoted to com.
18
-
of the policy under review.
Fann output increased in response to a net increase in variable factor usage. The increase in
farm output is also positively related to farm size, and falls within the range 235-2,868 bushels per
fann. Next, short-run fann profits from com PS increased. As evident, the producer's surplus change
varies positively with farm size. Small-sized farms record a producer surplus increase amounting to
about S200 (S250) for non-participants (participants), whereas large-sized farms record an increase
of about S2,445 (participating) and S3,281 (non-participating). Medium-sized farms come second
with a producer surplus increase of about S831 (S887) per farm for participants (non-participants).
The increase in short-run profits from com then led to an increase in full income. The changes in full
income are identical to those for short-run profits as indicated in Table 4 (i.e Y=1t+WT+V =>
dY=d 1t ; dT=dV=dW=O).
The increase in full income further led to an increase in leisure and money income. The
increase in leisure falls within the range 30-105 hours, and appears to generally vary positively with
class size. The increase in money income is also positively correlated with farm size. Now, since
both leisure and owned com hours have increased, with other farming hours and the time endowment
held unchanged, off-farm time declined drastically enough to just offset the combined increase in
leisure hours and owned com hours (Table 4). Because of the increase in leisure (or the net decrease
in total work time), the valuation of owned work labor increased and the laborer's surplus declined
as a result. Small-sized farm households registered a laborer's surplus reduction of about S180 per
farm, whereas medium-sized fanns experienced a decline ofabout S340. The large sized-participating
category recorded a laborer's surplus decrease of S568 per farm, whereas the large-nonparticipating
category registered a decrease of S283 per farm.
The producer's surplus increase should more that offset a laborer's surplus decrease so that,
19
-
in net, this policy would lead to a positive economic surplus change. The results in Table 4 confirm
this. The economic surplus change varied positively with farm size. The range was $21 through
$2,713 per farm. The results of this section indicate that the conventional analysis producer surplus
changes, at the exclusion of the laborer's surplus change, would generally lead to an overstatement
ofeconomic surplus change. Therefore, we now have adequate empirical evidence to suggest that,
following the convention of sole reliance on the producer surplus change, it would be quite likely to
overstate the improvement in household welfare, resulting from any particular market oriented
government program.
The implication of the distributional effects of the com program are very clear. It is evident
that the benefits to small-sized farms are relatively negligible. This class records an economic surplus
increase of only $21 ($65) per farm for non-participating (participating) households. The medium­
sized class saw an economic surplus increase of about $494 ($532) per farm for participating (non­
participating) households. Large-sized farms on the other hand saw the highest economic surplus
increase of $2, 163 ($2,713) per farm for non-participants (participants). These results would not
justify government support, if the overall objective ofgovernment intervention is to improve the well­
being of poor farm families. Therefore, the gains from government activity in the com market appear
to be misdirected. Hence, the objective of improving low farm incomes is clearly unrealized.
Concluding Remarks
The central objectives of this paper were to show that the producer's surplus is not an accurate
measure of farm-household welfare and that welfare consequences of farm policies vary across
producers. Empirical estimates of the laborer's surplus indicate that the economics of farm
households calls for the inclusion of the laborer's surplus in measuring welfare. Two observations
20
­
can be made concerning the measurement ofeconomic surplus. Ignoring the laborer's surplus would
(a) understate the magnitude of welfare currently enjoyed by farm households, and (b) overstate the
improvement in welfare resulting from the implementation of any market oriented policy. That is,
utilizing only the producer's surplus change to measure economic surplus change (induced by market
oriented programs), at the exclusion of the laborer's surplus change, would normally overstate
household welfare improvement. The reason is that, market oriented policies intended to increase
the producer surplus, would also lead to a reduction in the laborer's surplus, as households reallocate
their labor among competing alternatives. These results reinforce the argument that the household
can be viewed to allocate its resources in a way that maximizes the sum of the laborer's surplus and
the producer's surplus. Therefore, any time a market oriented (producer biased) policy is introduced,
the household would make a tradeoff between the two surpluses, in a way that would lead to a net
gain in household welfare. In particular, any policy geared towards increasing the producer's surplus
would lead to a reduction in the laborer's surplus, but the net effect would be a net increase in
economic surplus.
We also examined the distributional consequences of the com program among the various
groups of farms. The empirical results showed that the U.S. com program aids large-sized farms
relatively more than small-sized and medium-sized ones. Moreover, medium-sized farms have seen
greater economic surplus improvement, compared with small-sized farms. In point of fact, policy
induced improvement in the well-being of small-sized farms is so negligible that economic surplus
increases may be safely rounded to zero. A welfare improvement of$21 ($64) per farm was recorded
for small-nonparticipating (participating) households. It would indeed be not persuasive to argue that
small-sized farms would have been worse offin the absence of the program.
These results question whether program objectives are being realized in aiding the poor
21
­
fanner. Indeed, program benefits appear to be misdirected to households who do not seem to require
government support, at the expense of small and more needy farm households. Therefore, the
implications ofthese results are a clear indication that government involvement or activity in the U.S.
com market needs to be revised/revisited. If the objective is to sustain the poorest farm families in
farming, non-market oriented policies such as lump-sum transfers may be more effective since they
can target support to the most needy farm-households..
References
Bullock, D. S., "Objectives and Constraints of Government Policy: The Countercyclicity to
Agriculture", American Journal ofAgricultural Economics, 74(August 1992):617-629.
Floyd, 1. E., "The Effects of Farm Price Support on the Returns to Land and Labor in
Agriculture", Journal ofPolitical Economy, 73(April 1974): 148-58.
Gardner, B. L., "Efficient Redistribution Through Commodity Markets" American Journal of
Agricultural Economics, 64 (May 1983): 225-34
_ _ _ "Changing Economic Perspective of the Farm Problem" Journal ofEconomic Literature,
30(March 1992):62-101.
Gisser, M., "Price Support, Acreage Controls, and Efficient Redistribution", Journal of Political
Economy, 101(1991):585-611.
Huffman, W.E., "Interaction Between Farm and Nonfarm Labor Markets", American Journal
ofAgricultural Economics, 59(Dec. 1977): 1054-61.
_ _ "Agricultural Household Models: Surveys and Critique", in Hallberg, M. C., 1. L. Findeis
and D. A. Lass, eds, Multiple Job-Holding Fann Families. Ames: Iowa State University Press (1991)
Huffman, W. E., and M. Lange, "Off-farm Work Decisions of Husbands and Wives: Joint
Decision Making", The Review ofEconomics and Statistics, 71(August 1989):471-80.
Kawagoe, T., K., Otsuka, and Y. Hayami, "Induced Bias of Technical Change in Agriculture:
The United States and Japan, 1880-1980", Journal ofPolitical Economy, 94(June 1986):
523-44.
22
-
Lass, A. D., 1. L. Findeis, and M. C. Hallberg, "Factors Affecting the Supply ofOff-fann Labor: A
Review ofEmpirical Evidence" in Hallberg, M. C., 1. L. Findeis and D. A. Lass, eds, Multiple
Job-Holding Among Farm Families. Ames: Iowa State University Press (1991)
Nakajima, C., "Subsistence and Commercial Family Fanns: Some Theoretical Models of Subjective
Equilibrium", in Wharton, C. R, Jr. ed, Subsistence Agriculture and Economic Development,
Chicago: Aldine Publishing Company (1965).
_ _ _--', Subjective Equilibrium Theory of the Fann Household, New York: Elsevier (1986).
Nerlove, M., The Dynamics of Supply, Baltimore: John Hopkins University Press (1958).
Rosine, J., and P. Heimberger, "A Neoclassical Analysis of the U.S. Fann Sector, 1948-1970"
American Journal ofAgricultural Economics, 56(1974):717-29.
Seleka, T. B., Farm-Households and the Measurement of Economic Surplus: An Analysis of the
Distribution of U.S. Com Program Benefits Among Farm Households, Unpublished
Ph.D Dissertation, Cornell University, January 1996.
Shumway, C. R, H. Talpaz, and B. R Beattie, "The Factor Share Approach to production
Function "Estimation": Actual of Estimated Shares?", American Journal of Agricultural
Economics, 61(Aug. 1979): 561-564.
Singh, I., L. Squire, and 1. Strauss, eds, Agricultural Household Models: Extensions.
Applications. and Policy. Baltimore: John Hopkins University Press (1986a).
"The Basic Model: Theory, Empirical Results, and Policy Conclusions", in Singh, I.,
L. Squire, and 1. Strauss, eds, Agricultural Household Models: Extensions. Applications. and
Policy. Baltimore: John Hopkins University Press (1986b).
_ _ _----:>'
Sumner, D. A. "The Off-farm Labor Supply of Farmers", American Journal of Agricultural
Economics, 64(Aug. 1982):499-509.
Tyner, F. H. and L. G. Tweeten, "A Methodology for Estimating Production Parameters",
Journal ofFarm Economics, 47(1966): 1462-67.
USDA, Agricultural Statistics 1992. Washington, DC: United States Government Printing Office
(1992).
USDA, Agricultural Statistics 1993. Washington, DC: United States Government Printing Office
(1993).
Wallace, T. D. "Measures of Social Costs of Agricultural Programs", Journal of Farm Econ,
44(1962):580-94.
23
-
Appendix B
Economic Surplus Measurement
In stage I of optimization, the farm household determines optimal producer's surplus (n):
n··C'
IJ
=pqlJ
Q.. -W IJIJ
..r'"-PxX..IJ·
•
(B.I)
Note here that producer surplus is identical to restricted (shot-run) profits. IS
In stage 2, the optimal amount of the laborer's surplus is determined subject to optimal producer's
surplus. Therefore, following the first stage, the income constraint (equation I b) may be rewritten
as
I IJ.. =FI..+W..U
IJ
IJ IJ..
where
(B.2)
FI.IJ. =n~·IJ +R~+V
..IJ denotes fixed income and the second term where U..IJ =T..IJ -L..IJ denotes the
IJ
variable component of money income. Note that Uij=Tij-L ij represents total work time (off-farm
and farming time).
For a Cobb-Douglas utility function, the marginal valuation (cost) of family labor may be
expressed as
Q ..
=
IJ
UL(I L)
A..
I..
A..
I..
' = .!J!.--.!L = P 1J
IJ
U1(I,L) Ct IJ.. L..IJ Ct..IJ (T..-U..)
IJ IJ
(B.3)
Evidently, the marginal valuation offamily labor (equation B.3) is a function of work time (or leisure
time) and income. But since money income is also a function ofwork time (or leisure), we can utilize
equation B.2 to rewrite equation B.3 as
[FI.IJ. +W..(T..
IJ IJ -L..)]
IJ =.!J!.[FI..+W..
IJ
IJu..IJ ]
A ..
Q.. =.!J!.
IJ
Ct..
IJ
A ..
L..IJ
Ct..
IJ
T..-u..
IJ IJ
(B.4)
Equation B.4 indicates that the marginal valuation (cost) of family labor curve is upward
sloping with respect to owned work time and downward sloping with respect to leisure, as was drawn
IS Note that producer surplus is defined as the triangle-like area above the output supply curve
and below the product price line. This area is equivalent to restricted profits in the case where we
have concave production functions with respect to variable production factors. This should be
obvious because an(p q)/apq=Q(Pq)' From this result, it follows that fQ(P q)dPq= n(p q)'
Therefore, integrating the output supply curve should precisely define short-run farm profits.
-
in Figures 1 and 2. 16 The laborer's surplus can therefore be calculated as
_
= W..U..
IJ IJ
u..
T··
vFI..+WU..
_
vFI..+W..(T.. -L..)
-.!2!f
IJ
1JdU.. = W ..U .. -.!2!f IJ
IJ IJ IJ
a..
T..-U..
IJ
IJ IJ a..
L..
fl ..
IJO
fl ..
IJ
IJ
IJ
1Ji:..I)
(B.5)
= W..U.. + Pij[Iw.. T.. +FI..VI~ .. -lnT..)+w.h.. -iJl
IJ IJ a.. \.. IJ IJ
IJ}; IJ
IJ
IA IJ IJ/J
IJ
-
-
-
where U=T-L denotes the optimal amount of time devoted to work (farm and off-farm) and L is
the optimal amount of leisure time. 17 Total economic surplus for a representative household (ESij)'
This can be shown by partial differentiating equation B.4 with respect to leisure or total
work time to obtain
16
an = _1![FI +WT]= _ an <0
aL
a
L2
au
where the ij's have been dropped.
17
It is noteworthy here that
-
U
T
fFI +W(T-L) dL = -J![(WT+FI)(lnL-InT)+W(T-L)]
J!fFI+WU dU =J!
a o T-U
ai:
a
L
where the ij's have been dropped. Note that this expression denotes total valuation (cost) of
family labor devoted to work (all work). Therefore, the laborer's surplus is precisely calculated as
total wage income (farm and off-farm) minus total cost of owned labor. Recall that farm work is
assumed to be remunerated at the same wage rate as does off-farm work when calculating the
producer's surplus. The value unpaid (family) labor was, among other variable inputs, deducted
from farm revenues to compute the producer surplus. Thus, the producer's surplus does not
capture returns to family labor. The laborer's surplus calculation, therefore, is concerned with
measuring net returns to family labor put to work (farm and off-farm)..
25
-
Table I Distribution of off-farm income by sales class, 1986.
Off-farm income
Agricultural
Subsector
Percent of
sector's
farms
Average
$
Total
($000,000)
Percent of
sectors's
off-farm
income
All farms
100
20,212
44,708
46
100
Sales class
Less than $40,000
$40,00-$99,999
$100,000-$249,999
$250,000 or more
73
13
10
4
22,534
13,780
12,602
17,562
36,336
4,053
2,648
1,670
96
37
17
5
81
9
6
4
Source: Extracted from Ahearn and Lee (1991), Table 1.2
I
Percent total
cash income
from off- farm
sources
Table 2 Corn Production Costs Per Acre
Sales Class
Variable
less than $40,000 (Small)
$250,000 and over (Large)
Off-farm=Ycs
Off-farm = No
Off-farm=Yes
Off-farm = No
Off-farm = Yes
Off-farm = No
52,898.10
83.00
64,609.21
88.00
114,215.82
179.00
128,440.54
213.00
26,335.81
53.00
36,905.04
92.00
Average yield per acre
Gross value of production
85.69
194.43
80.05
186.48
112.27
256.51
105.85
242.96
119.80
281.04
114.51
266.92
IIired labor
Other Variable inputs
Total variable cash costs
1.10
113.73
114.83
2.54
106.16
108.70
3.92
122.67
126.59
3.17
127.60
130.77
12.63
138.34
150.97
14.28
156.94
157.97
General Farm overhead
Taxes and insurance
Total Interest
Total fixed cash costs
11.25
16.29
15.91
43.46
16.60
17.15
4.85
38.59
10.77
19.01
17.53
47.31
10.35
18.82
13.03
42.20
8.46
15.87
21.42
45.75
10.61
18.04
18.76
47.42
158.29
147.29
173.91
172.97
196.73
205.39
Gross value of production less cash costs
36.14
39.19
82.60
69.99
84.31
61.53
Capital replacement
21.99
22.04
23.26
25.95
29.13
33.83
Operating capital
Nonland capital
Net land rent
lJ npaid la bor
Total economic costs
3.12
10.17
52.78
41.09
271.53
2.96
10.11
54.51
49.35
281.41
3.44
9.33
67.13
26.18
285.73
3.56
10.19
57.26
29.34
286.23
4.11
10.12
64.17
14.35
297.19
4.30
11.23
61.22
13.58
310.79
Gross value production less economic costs
-77.10
-94.93
-29.22
-43.27
-16.15
-43.86
Total number of farms (expanded)
Total number of farms (sample)
Total cash costs
Source: USDA, Economic Research Service
~
$40,000-$249,999 (Medium)
I
Table 3 Parameter Estimates and Prices
Sales class
Variable
less than $40,000 (Small)
Production Parameters
A
'Y
w
()
t
Prices (dollars)
Wage rate
Price or other inputs
Producer Price (weighled AVG = 2.309)
$40,000-$249,999 (Medium)
$250,000 and over (Large)
Off-rarm= Yes
Off-rann=No
Off-rann= Yes
Off-rann = No
Off-rarm= Yes
Off-rann= No
2.692
0.168
0.615
0.216
3.623
2.977
0.210
0.570
0.220
3.544
3.199
0.102
0.635
0.262
2.812
2.678
0.126
0.651
0.223
3.490
2.225
0.053
0.712
0.235
3.252
2.425
0.099
0.684
0.217
3.608
3.725
1.000
2.269
3.584
1.000
2.330
4.267
1.000
2.276
4.560
1.000
2.295
5.129
1.000
2.346
5.138
1.000
2.331
0.550
0.450
0.185
0.815
0.607
0.393
0.473
0.527
0.836
0.164
0.789
0.211
Utility Function Parameters
Ct
{3
I
Table 4 Effects of the Corn Program (Changes in Exogenous Variables)
Sales class
Variable
less Ihan $40,000 (Small)
$250,000 and over (Large)
OIT-farm=Yes
Off-farm = No
Off-farm = Yes
OIT-farm = No
Off-farm = Yes
OIT-farm=No
52
711
1,255
54
529
1,679
23
76
2,014
11,372
110
2,593
1,968
119
143
9,937
1,430
216
7,713
1,598
317
266
14,083,420
235
15,189,250
530
60,553,070
883
113,376,800
2,868
75,542,210
2,528
93,285,190
Time endowmenl
Corn hours
Olher farm hours
OIT-farm hours
Leisure hours
0
1,255
0
-1,285
30
0
1,679
0
-1,725
46
0
1,372
0
-1,448
77
0
1,968
0
-2,070
103
0
1,430
0
-1,535
105
0
1,598
0
-1,699
100
Incomes per household
Properly Income
OfHarm wage income
Nel incomes for non-corn
Money income
Full income
0
-1,285
0
137
250
0
-1,725
0
38
204
0
-1,448
0
504
831
0
-2,070
0
420
887
0
-1,535
0
2,744
3,281
0
-1,699
0
1,930
2,445
250
13,218,070
-185
-9,796,450
0
0
65
3,421,622
204
13,189,160
-183
-11,839,800
0
831
94,934,130
-338
-38,548,700
0
0
494
56,385,420
887
113,968,800
-355
-45,608,700
3,281
86,407,800
-568
-14,966,700
0
0
2,713
714,411,300
2,445
90,246,630
-283
-10,428,500
0
0
2,163
7,981,813
Inpuls: F
X
Owned hours
Hired Labor (non-pari)
OUlpUI (bu): Per farm
Group Level
Welfare Changes
PS: Househo/d level
Group level
LS: Household
Group level
R: Household
Group level
ES: Household
Group level
I
$40,000-$249,999 (Medium)
°
21
1,349,313
°
0
532
68,390,080
t:
.­
r
-0
...
(L)
.­
::r:
...0
.c
ca
...J
J:
.......
­
~
CJ)
~
:.J
~.:'
...J
O'
...J
...-..
--------------;.---------------- 0 .c
::> 0
-- ()
..
C?
:,:
-­
0
t:
0
~
//
>
0
W
I
i
----'I!.......-·-~---O
«
E
a..
;' !
/ r>
OJ
aLUooul
~
en
0
.­
I/
/
-...a.
'r'"'
~
~
0
...
(L)
~
C)
.­
u..
-
J:
'­
o
~------'«--"7>~1r- ~_CJ~~--r-rr"'"7"--""'7'"":()"'-' r- j
.t:
....,
.­
.0
~
CJ)
-::JC.
'­
::J
CJ)
:.J
::>
i
·············
..J
"'
()
to;
-L--~----~o
«
awooul
aJ
o
.­
E
o
c:
0
0
__~---,o w
C\J
~
:J
C)
.­
u..
-
Ea'i
v\
1M
(a)
8a
c
,
.,~~",~,
,
---i._
i
~
: ·dL~d
,
,
,,
,
,
,'
.
,
o
1 I'
l' 1
, (b)
$
o
1I' l' 1
~
Ig
" \ 19
----)V
T
M'
-----. h
T
Figure 3 Effects 01 a Price Support
~
I
Appendix A
Data Used in Model Calibration
Table A.I Corn Production Costs Per Acre
Sales Class
Variable
less Ihan $40,000 (Small)
$250,000 and over (Large)
Off-farm=Yes
Off-farm = No
Off-fann= Yes
Off-farm=No
Off-farm = Yes
Off-farm = No
52,898.10
83.00
64,609.21
88.00
114,215.82
179.00
128,440.54
213.00
26,335.81
53.00
36,905.04
92.00
Average yield per acre
Gross value of production
85.69
194.43
80.05
186.48
112.27
256.51
105.85
242.96
119.80
281.04
114.51
266.92
lIired labor
Olher Variable inputs
Tolal variable cash coslS
1.10
113.73
114.83
2.54
106.16
108.70
3.92
122.67
126.59
3.17
127.60
130.77
12.63
138.34
150.97
14.28
156.94
157.97
General Fann overhead
Taxes and insurance
TOlal Inleresl
TOlal fixed cash costs
11.25
16.29
15.91
43.46
16.60
17.15
4.85
38.59
10.77
19.01
17.53
47.31
10.35
18.82
13.03
42.20
8.46
15.87
21.42
45.75
10.61
18.04
18.76
47.42
158.29
147.29
173.91
172.97
196.73
205.39
Tolal number of farms (expanded)
TOlal number of farms (sample)
TOIal cash cosls
I
$40,000-$249,999 (Medium)
Table A.I (Continued) Corn Production Costs Per Acre
Sales Class
Variable
less than $40,000 (Small)
$250,000 and over (Large)
Off-farm = Yes
Off-farm = No
Off-farm = Yes
Off-fann=No
Off-fann=Yes
Off-farm = No
Gross value of production less cash costs
36.14
39.19
82.60
69.99
84.31
61.53
Capital replacement
21.99
22.04
23.26
25.95
29.13
33.83
Operating capital
Nonland capital
Net land rent
Unpaid labor
Total economic costs
3.12
10.17
52.78
41.09
271.53
2.96
10.11
54.51
49.35
281.41
3.44
9.33
67.13
26.18
285.73
3.56
10.19
57.26
29.34
286.23
4.11
10.12
64.17
14.35
297.19
4.30
11.23
61.22
13.58
310.79
Gross value production less economic costs
-77.10
-94.93
-29.22
-43.27
-16.15
-43.86
Source: USDA, Economic Research Service
I
$40,000-$249,999 (Medium)
Table A.2 Other Pertinent Data
Sales Class
Variable
Small
Large
On:farm= Yes
Off-farm = No
Off-limn = Yes
on: farm = No
Off-farm = Yes
On:farm=No
1,575
28
1,907
23
2,663
30
2,938
28
2,880
33
3,123
25
Hours of other household members per year
% hours for corn
714
17
647
II
893
15
1,237
12
1,889
26
1,747
14
Other sources of farm income
Customs work for others
Grazing of livestock
Cooperative patronage dividends
Sales of farm machinery and vehicles
Insurance indemnity and payments
Hunting, fishing and outdoor recreation
708
0
72
189
84
0
51
0
140
19
151
0
1,312
46
452
436
698
5
1,049
II
691
214
684
3
4,918
1,918
1,132
310
1,574
0
4,653
390
868
2,200
1,024
27
22,500
500
0
500
500
0
1,750
500
1,750
1,750
12,500
500
0
500
500
0
500
0
500
500
17,500
500
0
1,750
1,750
0
500
0
1,750
500
I
51
4
0
37
2
12
140
12
22
129
12
102
413
41
142
246
32
Hours of operator per year
% hours for corn
Off-farm income
Cash wages and salaries
Net cash income from on:larm husiness
Net cash income Ihml another farm or ranch
Interest and dividends
Other off-farm income
Acres planted to corn
Inigated acres
Non-irrigated acres
Set-aside acres
Source: USDA, Economic Research Service
I
Medium
Table A.3 Income Statement for Farm Business, 1991 FCRS, Corn Version
Sales Class
Variahle
less Ihan $40,000 (Small)
Total number of farms (expanded)
Total numher of farms (sample)
Tolal acres operated
Corn acres planted
Other crop acres planted
Average cash rent per acre
Average cash rent and AUM expense
Shared rent estimale incl. liveslock
Gross Cash Income
·
·
·
·
I
Livestock sales
Crop sales
Government paymenls
Other larm-relaled in<:ome
$40,000-$249,999 (Medium)
$250,000 and over (Large)
Off-tarm= Yes
Off-farm = No
Off-farm = Yes
Off-farm = No
Off-farm = Yes
52,898
83
64,609
88
114,216
179
128,441
213
26,336
53
36,905
92
229
51
18
55
2,232
2,293
214
38
28
30
588
1,028
471
157
48
60
8,641
14,425
526
156
45
38
5,642
16,142
1,855
519
50
73
27,764
30,414
1,764
398
148
46
26,990
28,797
Oll~farm=
No
20,351
20,842
102,971
110,097
443,771
592,418
7,386
10,492
1,352
1,120
9,875
8,300
1,077
1,590
43,529
47,898
5,681
5,863
53,853
45,026
5,581
5,637
207,108
199,968
20,049
16,646
335,670
214,753
19,921
22,074
Table A.3 (Continued) Income Statement for Farm Business, 1991 FCRS, Corn Version
OIT-farm=¥es
On~larm=No
(>1'1'- farm = ¥ es
Off-farm = No
OIT-farm=¥es
OIT-larm=No
22,808
19,529
81,360
82,518
352,387
454,926
15,838
1,134
1,716
179
1,445
3,776
390
1,600
2,631
561
672
1,433
15,552
910
1,959
114
1,154
4,018
744
1,453
2,456
563
790
1,038
58,709
5,853
10,991
333
4,679
13,257
2,562
4,534
7,144
1,827
2,316
3,666
62,259
5,522
11,456
537
4,489
13,405
3,558
4,783
7,994
2,455
2,498
3,746
269,540
59,370
37,767
1,170
18,139
54,379
25,056
19,392
25,477
7,734
6,696
9,503
377,041
124,265
72,705
1,849
14,166
48,708
32,948
16,419
22,541
12,143
9,631
14,137
-Fixed
· Real estate, property taxes
· Interest
· Insurance premium
· Rent and lease payments
6,967
1,125
2,614
940
2,288
3,977
1,576
916
875
610
22,651
2,047
8,806
2,670
9,128
20,258
2,763
8,394
2,947
6,155
82,847
9,198
35,035
7,463
31,151
77,885
8,186
26,898
13,058
29,743
Equal: Net cash farm income
-2,453
1,313
21,611
27,580
91,384
137,492
Less: Depreciation
Labor, non-cash benefits
3,428
2
2,325
58
10,466
68
11,480
270
28,571
1,405
30,648
2,064
Plus: Value of inventory change
Nonmoney income
3,917
2,495
2,049
3,084
2,285
2,667
2,344
2,632
27,222
4,898
11,164
5,053
529
4,063
16,027
20,805
93,527
121,002
Variable
Less: Cash Expenses
- Variable
· l.ivestock purchases
· Feed
· Olher livestock expenses
· Seed and plants
· Fertilizer and chemicals
· Labor
· Fuel and oils
· Repairs and maintenance
· Machine-hire and custom work
· Utilities
· Other variable expenses
Equals: Net farm income
Source: USDA, Economic Research Service
I
OTHER AoRoMoEo WORKING PAPERS
No. 96-05
Economic Growth, Trade and the
Environment: An Econometric
Evaluation of the Environmental
Kuznets Curve
Vivek Suri
Duane Chapman
No. 96-06
Property Taxation and Participation
in Federal Easement Programs:
Evidence from the 1992 pilot
Wetlands Reserve Program
Gregory L. Poe
No. 96-07
Commodity Futures Prices as
Forecasts
William G. Tomek
No. 96-08
Old-Growth Forest and Jobs
Jon M. Conrad
No. 96-09
The Economic Threshold With a
Stochastic Pest Population: An
Application to the European Red
Mite
Jean-Daniel Saphores
Jon M. Conrad
No. 96-10
Competitive Electricity Markets in
New York State: Empirical Impacts
of Industry Restructuring
Robert Ethier
Timothy D. Mount
No. 96-11
Do Participants in Well Water
Testing Programs Update Their
Exposure and Health Risk
Perceptions?
Gregory Poe
Harold van Es
Maxine Duroe
Christopher Crossman
Timothy VandenBerg
Richard Bishop
No. 96-12
A Hedonic Approach to Estimating
Operation and Maintenance Costs for
New York Municipal Water Systems
Todd M. Schmit
Richard N. Boisvert
No. 96-13
The Productivity of Dairy Farms
Measured by Non-Parametric
Malmquist Indices
Loren W. Tauer
-